AVALANCHE. inside. Volume 69 Summer Presenting Partners of the Avalanche News. government. correspondence. partners. public programs.

Transcription

1 AVALANCHE news Volume 69 Summer 2004 Presenting Partners of the Avalanche News inside government correspondence partners public programs industry upcoming events education history new products Published by: Canadian Avalanche Association Box 2759 Revelstoke, BC V0E 2S0 (250)

2 Review of Glide Processes and Glide Avalanche Release Alan Jones Abstract Glide, the process whereby a snowpack on a slope moves along the ground interface, is an important process that often leads to the formation and release of glide avalanches. Glide avalanches often mobilize large volumes of snow, and can have major destructive potential. A review of gliding concepts, glide mechanics, glide avalanche trigger mechanisms, factors affecting glide rates, methods for forecasting, and artificial release (control) of glide avalanches is presented. The understanding of glide processes and glide avalanche formation and release has increased significantly in the last 40 years, with the development of numerous models based on constitutive equations and mechanics of materials. Research has shown that the amount of free water at the interface between the snowpack and the ground is likely the most important factor affecting glide rate and acceleration, which are in turn understood to be important indicators of natural glide avalanche release. The amount of free water at the snow/ground interface is controlled by precipitation from rain-on-snow events, melting due to solar radiation, and melting of the snowpack from higher air temperatures. Progress has been made over the last 40 years to better forecast glide avalanche release, while methods for effectively releasing glide avalanches using artificial means are somewhat limited at present. 1. Introduction In the classification of avalanches, distinctions are typically made between dry snow avalanches and wet snow avalanches. Avalanches consisting of dry snow are responsible for most of the damage and fatalities from avalanches (McClung and Schaerer, 1993). However, the destructive potential of wet snow avalanches has also been acknowledged as important, particularly in locations where human lives or structures are threatened (Wilson et al., 1996). A glide avalanche (Figure 1) is a unique type of avalanche, that may loosely be classified as a wet avalanche (e.g. McClung and Schaerer, 1993, p. 83), and is particularly dangerous due to the potentially large volume of snow involved and difficulty to forecast release. Snow glide is associated with glide avalanches, and therefore an understanding of glide processes is considered critical to understanding how glide avalanches develop and often lead to full-depth releases of the snowpack. Figure 1. Slab avalanche released by gliding snow cover (Photo credit: R. Perla) In this paper, the processes and mechanisms associated with glide avalanche formation and release are discussed. This is followed by a discussion of the difficulties involved with forecasting for glide avalanches, and some of the forecasting methods that are available to the avalanche forecaster. 2. Background One of the earliest mentions of the glide process was by Seligman (1936), where he proposed the name snow rift to describe the development of a crack in a snow slope, now commonly known as a glide crack. The first period of extensive of glide processes and glide avalanches was conducted in Switzerland in the late 1930 s through the 1940 s, predominantly by the Swiss ers Haefeli, Bader and Bucher (Bader et al., 1939). Their work included laboratory tests to study glide velocity, and descriptions of the relationship of glide to interface temperature and snow viscosity, shear stress conditions and differences between wet and dry gliding. Important conclusions of from this period include: (1) gliding velocity increases linearly with shear stress and reaches a constant value; (2) gliding velocity depends on the stress condition and the interface temperature; (3) the friction coefficient decreases from a high value for dry gliding to a low value for wet gliding; (4) when a critical gliding velocity is achieved, the process becomes dynamic and avalanche formation results; and (5) glide velocity, similar to creep velocity, is determined by shear stress and snow viscosity (Bader et al., 1939). They summarized the most important factors that have an effect on glide as: slope exposure; degree of surface roughness; temperature of the ground surface; and the thickness and properties of the snow cover. They proposed that the primary action of these factors is their effect on the frictional conditions of the slide surface. 53

3 Between the mid 1950 s and late 1960 s, In der Gand and Zupanièiè conducted on glide avalanches for the Federal Institute for Snow and Avalanche Research in Davos, Switzerland. A summary of field investigations in the Swiss Alps was provided by In der Gand and Zupanièiè (1966), describing the influence of ground surface roughness, terrain shape and snow characteristics on glide avalanches. A method for measuring glide velocity was presented, including the introduction of the use of glide shoes to measure glide on avalanche slopes, and a glide velocity equation was developed from field measurements and glide theory. During the 1970 s and 1980 s, much of the theoretical understanding of glide and glide avalanches was developed in Canada, Switzerland and Japan. McClung (1980; 1981; 1987a; and 1987b), developed various aspects of glide avalanche theory, including: constitutive equations that relate shear stress to glide velocity (McClung, 1980); constitutive equations relating tangential drag on the snowpack to slip velocity (McClung, 1981); a model for glide crack initiation (McClung 1987); and the effects of free water at the snow-earth interface on glide velocity (McClung and Clarke, 1987). Lackinger (1987) conducted field experiments in Switzerland on glide avalanches to develop a theory for glide crack formation and the mechanism of release for glide avalanches. Results were presented relating glide rates to the climate prior to full-depth avalanche release. Concurrent with in Canada and Switzerland, Japanese ers were conducting field and laboratory studies to describe glide mechanisms on slopes covered with bamboo bushes in Japan. Most notably, Endo and Akitaya (1978) and Endo (1983; 1985) described the anchoring effect of vegetation on a slope for full-depth avalanche release, and provided further insight into glide crack formation (Endo, 1983; Endo and Akitaya, 1978). The most recent glide avalanche reported in the literature is for field studies conducted on slopes adjacent to the Coquihalla Highway, in the Cascade Mountains of British Columbia. Results from these studies describe the temporal and spatial dependence of gliding (McClung et. al, 1994), and expand on current understanding of snow glide rates and their relationship to climatic conditions (Clarke and McClung, 1999). 3. Concepts of snow gliding On a slope, snow is affected by three general deformation processes: settlement (perpendicular to snow surface); internal shear deformation (parallel to snow surface); and glide (Figure 2). The term creep is used to describe the resultant of settlement and internal shear deformation. The third component, snow glide, is the translational slip of the entire snowpack along an interface, typically either the ground surface or a thin water/ice layer (McClung and Schaerer, 1993). Glide occurs on both dry and wet snowpacks, but is typically negligible on dry snowpacks because dry snow has a high degree of friction along the ground interface (McClung and Schaerer, 1993). In wet snowpacks, snow gliding is a more important process that can result in full-depth avalanche release with large impact forces, as well as deformation of structures located along a slope where glide is occurring. Previous field studies have shown that there are three basic prerequisites that must be met before gliding can occur (Clarke and McClung, 1999): Figure 2. Components for snowpack deformation (after McClung and Schaerer, 1993) (1) the interface must be smooth (e.g. bare rock or grassy vegetation); (2) the temperature at the snow/ground interface must be at 0º Celsius, guaranteeing the presence of free water at the interface; and (3) the slope angle must be at least 15º for roughness typical of alpine ground cover. Surfaces that provide a smooth snow/ground interface, such as polished rock slabs or grassy slopes, are associated with the fastest glide rates. (McClung and Clarke, 1999). For glide to occur on rock, free water must reach the rock/snow interface. Glide on grass-covered slopes can occur without the presence of free water, but is greatly enhanced by the presence of free water (McClung and Clarke, 1999). For vegetative surfaces with high roughness (e.g. a harvested forest area, slide alder), glide is typically not observed. 54 Studies have shown that the rate of gliding is very sensitive to the amount of free water present at the snow/ground interface (e.g. Clarke and McClung, 1999; Lackinger, 1987). Sources of free water to the interface include: (1) rainfall; (2) melt at the interface resulting from stored summer heat; (3) snowpack melt by solar (short wave) radiation; and (4) melt from geothermal hot spots and groundwater outflows (McClung and Clarke, 1987). The first three sources are typically considered most important, while the effects of geothermal hot spots and groundwater outflows are seldom observed and are of limited interest.

4 Glide only occurs when the slope angle of the ground is at least 15º. Above this angle, the downslope portion of the gravitational force acting on the snowpack is larger than the combined frictional forces from the snow/ground interface and the internal frictional forces within the snowpack. According to the model developed by McClung and Clarke (1987), free water at the interface has two principal effects on glide mechanics: it promotes separation of the snowpack from the ground at the interface; and it decreases the snow viscosity. By increasing the separation of the snowpack from the surface, irregularities on the ground surface tend to be drowned out, allowing the snow to move downslope unimpeded (McClung and Clarke, 1987). Free water available at the surface influences the stiffness of the snow slab because the slab viscosity decreases with increasing water content, making movement of the snowpack over ground roughness features easier (McClung and Clarke, 1987). An increase in the snow/ground separation and a reduction in snow viscosity can sometimes combine to reduce the friction over a critical region of the base of the snowpack. The area of this region has been estimated to be 10 slab thicknesses long and wide (McClung and Shaerer, 1993), although there are no field studies to support this hypothesis. When friction is reduced in a downslope part of the slab, stress is transferred to the upslope portion of the slab, above the area of reduced friction. Tensile stress results in the upslope part of the slab, often resulting in formation of a tensile fracture, or glide crack (Figure 3). This crack initiates at the snow/ground interface and propagates upwards toward the snow surface, perpendicular to the interface (McClung, 1987). Figure 3. Tensile crack formation in the snowpack (Photo credit: Bruce Jamieson) Field observations show that glide crack formation is a prerequisite for full-depth glide avalanches, but that glide avalanche release does not always immediately follow glide crack formation (Clarke and McClung, 1999). The time between glide crack formation and glide avalanche release can typically range from a few hours up to several weeks (Lackinger, 1987), and occasionally up to months (McClung and Shaerer, 1993). The locations where glide cracks tend to form are highly dependent on the ground topography (Lackinger, 1987). The most common locations for glide crack formation include convex rolls, where tensile stresses are concentrated; on slopes where the ground bed surface roughness changes (e.g. transition between rock slab and grassy slope); and below steps in rock (McClung and Schaerer, 1993). Glide crack formation is also dependent on ground roughness and slope angle (McClung, 1987). In der Gand and Zupanièiè (1966) suggest a slope angle of 30º as a lower limit for glide crack formation, applicable to grass covered slopes or rock slabs. Areas with higher surface roughness may require slope angles greater than 40º for glide crack formation (McClung, 1987). 4. Snow glide mechanics Swiss, Canadian and Japanese ers have developed several glide models. The most recent, and perhaps most comprehensive theory was developed by McClung and Clarke (1987). This theory expands on a previous theory developed by McClung in the early 1980 s (McClung, 1980; McClung, 1981) and is discussed in the following section. McClung and Clarke (1987) developed a model that assumes that the water within the snowpack and at the snow/ground interface is the critical parameter that determines glide velocity and glide avalanche release. Their model provides a constitutive equation that relates the local basal stress, ô, at the snow/ground interface to the glide velocity, U 0, as: µ U 0 τ = 2(1 ν ) D, where D * is the stagnation depth, ì is the shear viscosity and õ is the viscous Poisson Ratio of the snow above the snow/ ground interface. The parameter D* is a geometric construct that is a function of the snow/ground interface geometry and water distribution at the interface (Figure4). If the glide velocity, U 0, is zero either due to high surface roughness or a subfreezing interface temperature, the stagnation depth, D * is also zero. The shear viscosity and viscous Poisson Ratio are functions of the changing density of the snowpack due to shear and compressive deformations, and the water content of the snowpack. Thus, if a significant amount of free water is present at the interface, the shear viscosity may be reduced, resulting in increasing glide rate. 55

5 McClung and Clarke (1987) emphasize the effect of water on the interface geometry, rather than the effects of varying shear viscosity and viscous Poisson Ratio with varying water content. Further studies by McClung et al. (1994) found that measurements of the characteristics of the snow at the interface with the ground showed only minor variation through the winter, while glide rates fluctuated substantially through the winter. Thus, they conclude that the effects of water on partial separation of the snowpack from the glide interface and in filling of irregularities in the ground has a greater affect on glide velocity than varying snow properties. 5. Glide avalanche trigger mechanisms Full-depth glide avalanche release is understood to be triggered by three general events: (1) loading by new snow; (2) rain-on-snow events; and (3) snowmelt. Snowmelt can either be from incoming solar (short wave) radiation, or by warming air temperatures that result in higher snowpack temperatures (Clarke and McClung, 1999). Lackinger (1987) found that glide avalanches were released more often after high air temperatures persisted for several days than after periods of rain. Release due to loading by fresh snow was only observed in one case during a period of record snowfall. Clarke and McClung (1999) found that most full-depth avalanches were triggered after the input of water from rain or snowmelt. The results of interpreted trigger mechanisms of 21 avalanches by Lackinger and 104 avalanches by Clarke and McClung are summarized in Figure 5. The difference between the number of rain-on-snow events between Lackinger and McClung and Clarke is likely because Lackinger observed rain-on snow events concurrent with the associated melting, and interpreted the snowmelt as being the more critical trigger mechanism. There may also be significant climatic differences between the Swiss Alps and Canadian Cascade mountain ranges that could partially account for these differences. Figure 5. Full-depth glide avalanche trigger mechanisms (source data from Lackinger, 1987, and Clarke and McClung, 1999) Figure 4. Schematic of creep and glide velocity components for the snowpack showing geometrical construction of D*, the stagnation depth (from McClung and Clarke 1987, with modifications) Release of glide avalanches on days with cold temperatures, although uncommon, was observed by both Lackinger and Clarke and McClung. The latter explain this phenomenon by the presence of melt-water at the snow/ground interface that is unaffected by low air temperatures due to snowpack acting as an insulator. Exposed rock surfaces near the upper part of the avalanche release zone may be warmed by solar radiation on cold days and cause additional snow melting, and this may be sufficient to trigger a glide avalanche. An alternate explanation was provided by McClung and Schaerer (1993), whereby diurnal changes in temperature cause the surface snow to contract as it freezes, due to the low water content (as would be expected near the surface). This may result in increased tensile stress in the crown area and fracture formation and lead to avalanche release. This may only apply to thin, wet slabs, making this an uncommon occurrence. Lackinger (1987) proposed that nighttime cooling makes the snow surface and flanks of the tension cracks freeze, changing the stress pattern within the snowpack. This may result in the damming up of meltwater at the snow/surface interface, and possibly lead to glide avalanche release. This same process may be responsible for avalanches triggered during cold periods Factors affecting glide rates It is generally accepted that the supply of free water to the snow/ground interface is the principal mechanism controlling snow glide and glide avalanche release (Clarke and McClung, 1999). This supply of free water is critical in determining the glide rate. In der Gand and Zupanièiè (1966) surmised that there is a critical value of gliding velocity that can be used to forecast glide avalanche release. Much of the subsequent has focused on investigating the relationship between rapid rates of glide, climatic conditions and the release of glide avalanches. Most ers have been confounded by this relationship, due to the difficulty in obtaining accurate measurements of glide avalanches and time lags between climatic effects and avalanche release. Clarke and McClung (1999) found that there was no clear direct relationship

6 between glide rates and full-depth avalanche occurrence for their data. They proposed that full-depth glide avalanche releases may best correlate with periods of increased glide acceleration, rather than increased glide rates. This has yet to be investigated, and therefore, glide rate is still used as the most likely indicator of glide avalanche release. The following sections describe the relationship between glide rates and temporal, spatial and climatic variations. 6.1 Seasonal variation Glide rates vary at any particular site throughout the winter season, and from year to year. Lackinger (1987) proposed that glide avalanches might be expected to occur in a season when gliding motion begins early in the season, particularly in seasons with early, heavy snowfall. McClung et al. (1994) noted higher glide rates and glide rate fluctuations in early season and attributed it to: (1) summer heat stored in the rock causing melting at the interface and (2) the early season snow is of low density, has a low shear viscosity and therefore deforms faster at the snow/rock interface. They also noted higher glide rates and fluctuations late in the season, corresponding to spring conditions. At this time of year, higher glide rates can be attributed to significant melting, warmer temperatures and higher likelihood of rain-on-snow events (Clarke and McClung, 1999). Figure 6 shows Figure 6. Mean monthly glide rates from Coquihalla Highway, Cascade Mountains, British Columbia (source data from Clarke and McClung, 1999) the mean monthly glide rate for 4 glide avalanche locations over a two-year period. Although there is significant variation, the general trend shows highest glide rates at the beginning of the season (i.e. November and December), lowest glide rates mid-winter (i.e. January to late February), and slightly increasing glide rates toward the end of the season (i.e. March to April). In der Gand and Zupanièiè (1966) noted a similar seasonal trend at a number of Swiss sites, with glide rates slowing down in the middle of January, and increasing in the middle of February, with increasing snowpack temperatures. Typical glide velocities are in the range of 1 to 100 mm/day (McClung and Schaerer, 1993), with mean monthly glide rates typically ranging between 5 and 20 mm/day (Figure 6). Data from the Coquihalla Highway (McClung et. al, 1994) indicated a similar seasonal pattern of gliding rates on a year-to-year basis. This was attributed to the similar snowpack and weather patterns for the two seasons presented. It is likely that comparison of glide rates for a given site on a long term basis will show larger variations in glide rates, although there is no to support this hypothesis. Figure 7 shows the mean seasonal (November through April) glide rates from 9 study sites on the Coquihalla Highway for the two-year period. The overall trend shows similar seasonal glide rates for each study site for the two-year period, with three sites (Shoe 4, Shoe 5 and Gauge 5) showing a large change between years. Figure 7. Mean seasonal (Nov.-Apr.) glide rates from Coquihalla Highway, Cascade Mountains, British Columbia (source data from Clarke and McClung, 1999 and McClung et. al., 1994) 6.2 Spatial variation Glide rates on a slope are highly variable, both downslope and across the slope. McClung et al. (1994) note that glide rates increased in the downhill direction. Endo and Akitaya (1978) found that after a glide crack formed, the glide velocity decreases on the slope above the crack, but increases on the slope below the crack. Clarke and McClung (1999) found that glide rates vary across the slope of a glide avalanche, making it difficult to specify a threshold glide rate for avalanche release. Variations in glide rates are thought to arise from variable shear resistance at the snow/ground interface that result from natural features of the snowpack and topography (e.g. slope inclination, roughness, snow depth, viscosity) (Lackinger, 1987). With variable shear resistance across and up the slope, glide velocity gradients occur in the snowpack, which give rise to stress gradients. The snowpack deforms in response to these stress gradients, but when the tensile strength of the snowpack is exceeded, glide cracks form and subsequent glide avalanche release often occurs. 57

7 Due to the high spatial variation in glide rates between adjacent slopes, monitoring glide rates at sites away from avalanche initiation zones is of limited use for forecasting purposes (Clarke and McClung, 1999). 6.3 Diurnal variation There is no conclusive evidence to relate glide avalanche release to diurnal variations. Lackinger (1987) found that avalanche release often occurred during the night. They surmise that nighttime cooling makes the snow surface and lateral edges of the glide cracks freeze, changing the stress distribution and causing free water to dam up along the ground/snow interface. In contrast to their findings, McClung et al. (1994) found that gliding was generally faster during the day for one year of data, and that there was no difference between nighttime and daytime glide rates for a second year of data. Further analysis (Clarke and McClung, 1999) for the same study site leads to the conclusion that there is no significant difference between nighttime and daytime glide rates. 6.4 Air temperature variation Most studies of glide have confirmed that air temperature strongly influences glide rate (e.g. In der Gand and Zupanièiè, 1966; Akitaya and Endo, 1984; Lackinger, 1987; Clarke and McClung, 1999). However, this relationship is complex and multivariate in nature, making it difficult to use air temperature as a forecasting tool for glide avalanches. Other factors that complicate this relationship include loading by precipitation (snow and rain), snowmelt by solar radiation, direct water inputs to the interface from rainfall (McClung et al., 1994). Cross-correlation analyses between glide velocity and air temperature by Clarke and McClung (1999) showed that glide rate at their sites correlates positively with air temperature, with lag times of 12 to 24 hours. Their results also show that the correlation between glide velocity and air temperature is higher during snowmelt-triggered avalanches than during rain-on-snow events. 7. Forecasting glide avalanches Numerous methods have been proposed over the years to try to forecast the release of glide avalanches, but the level of success has been limited. Consequently, forecasting glide avalanches remains a difficult process and relies heavily on the experience of the forecaster. Lackinger (1987) presented the results of a study where two measures were used to forecast: (1) measuring the glide velocity on the active glide avalanche; and (2) measuring the micro seismic and acoustic emissions (MSE) in the area of the glide avalanche. The concept behind measuring MSE is that during periods of increased glide velocities and acceleration, which are believed to precede full-depth avalanche release, increased amplitudes and frequencies of MSE can be measured, allowing for prediction of imminent avalanche release. Using MSE measurements, Lackinger (1987) was able to predict a small glide avalanche 3 hours before release due to an increase in the frequency and amplitude of MSE signals. However, these readings are complicated by the fact that signals originate both from the micro-fracturing of the snow, and an increase of gliding noise from the snowpack moving along the ground surface. French ers also tried similar measurements (Lackinger, 1987), but found that they measured lots of extraneous signals from things such as helicopters and earthquakes. Distinguishing noise signals from the snowpack (e.g. due to creep) from noise signals from those generated at the snow/ground interface complicates signal differentiation. A second method field tested by Rice et al. (1996) at the Alta ski area, Utah, involves the use of sensors that are designed to measure snow creep and glide in avalanche start zones. These sensors included an accelerometer and temperature sensor. One sensor was placed at the ground surface, while the other two were place one and two metres above the ground, respectively, allowing measurement of differential creep and glide movements. Using these instruments, there was some success in measuring the elastic response of the snowpack to nearby explosive use, and measurement of increasing creep rates prior to avalanching. These sensors seemed best suited for measuring creep, and their utility for measuring glide was limited, based on their results. Wilson et al. (1996) tried similar field experiments near Revelstoke, British Columbia, using a sprung probe in the avalanche start zone that tilts as the snow moves downslope. The probe is pushed over by the snow until an avalanche occurs, at which point it returns to the upright position ready to monitor the next avalanche release. Again, the probe uses an accelerometer and temperature sensor for monitoring purposes. Snowpack movement was monitored by measuring the tilt of the probe, and transmitting signals via radio communication to a forecasting centre. Due to the large forces resulting from creep and glide, some of the probes were damaged during trials. Results from this study show that the probes can monitor snow movement prior to avalanche release with some reliability. Limitations include: that if there are no glide avalanche releases, the probes don t return to their upright position; this system doesn t differentiate between glide avalanches and normal slab avalanches; and this method does not differentiate between glide and creep. 58

8 All the three methods described above are limited by the fact that they have difficulty differentiating between glide and creep processes in the start zone. The one method developed to date that allows discrete measurement of glide rates is the use of glide shoes. Glide shoes were first developed by In der Gand and M. Zupanièiè (1966), and subsequent installations of glide shoes have been based primarily on this pioneering work. Glide shoes are flat-bottomed steel boxes with inner baffles that are placed on the glide surface before snowfall (Clarke and McClung, 1999). The glide shoes are open at the top so that snow can fall into them. The shoes are connected to a potentiometer with a cord, which measures displacement of the glide shoe. This information is typically recorded on a data logger and downloaded via a modem to a forecasting centre. Use of glide shoes allows the measurement of glide displacements, and the corresponding velocity and acceleration due to the movement of the snowpack over the interface. Perhaps the most reliable method of forecasting glide avalanches uses a combination of one of the methods described above to measure glide rate and acceleration, combined with the monitoring of meteorological conditions. Meteorological conditions have proved to be critical in determining the release of full-depth glide avalanches (e.g. Clarke and McClung, 1999; Lackinger, 1987), particularly conditions that produce free water. Meteorological variables that are likely most important for forecasting glide avalanches include: precipitation (particularly rain from rain-on-snow events); air temperature (and the related snowpack temperature); and solar radiation. 8. Control of glide avalanches Very little has been conducted to date on the use of artificial control methods for glide avalanches. There is a lack of published articles in this field which is likely related to the limited success rate of control methods. Lackinger (1987) makes mention of the Schmalzberg avalanche that occurred in Switzerland in 1974, with a death toll of 12 people. Attempts were made to release the avalanche by using explosives along a glide crack, but were unsuccessful. An avalanche measuring 2 m thick by 100 m wide finally released due to melting water on the slope and recent snowfall, burying a ski tow and killing 12 people. In contrast to dry slab avalanches, increasing the load on the snowpack on a glide avalanche does not generally trigger glide avalanches. This is partly due to the high frictional forces from the mass of the glide avalanche slab against the ground. More free water at this interface serves to reduce this friction, making avalanche release more likely. Because of this high friction and the large mass of snow, control of glide avalanches by explosives is generally unsuccessful. Perhaps a better method that has been proposed (B. Jamieson, personal communication) would be to introduce additional water into the avalanche start zone. This could entail introducing water into a glide crack from the air (e.g. helicopter water bombing), pumping water into the start zone, or using heat tape or other heat conductors at the snow/ground interface to generate additional free water at the interface. Such methods may have been tried, but little information on control methods exists in the literature. 9. Conclusions Understanding of snow glide processes and glide avalanche formation has increased significantly in the last 40 years. Before the 1960 s, glide processes and glide avalanches were a poorly understood phenomena that occasionally had significant consequences, resulting in impacts on human lives and structures. Development of an understanding of the mechanics associated with glide has provided ers with some important tools for understanding how these avalanches form and when they may release. Important developments include the understanding that free water at the interface between the snow and the ground surface is likely the most important factor in the glide process. The amount of free water available is the principal factor that controls the glide rate and acceleration of the snowpack over the surface, which in turn affects when a glide avalanche will occur. Factors that contribute significantly to the amount of free water available at the snow/ground interface include rain-on-snow events, and melting due to solar radiation and warming from higher air and snowpack temperatures. The knowledge of methods for forecasting and artificial release of glide avalanches is relatively limited when compared to methods available for dry slab avalanches. Some progress has been made in recent years for forecasting when glide avalanches will occur, but ways of controlling these avalanches are limited at present. Additional and field studies could help forecasters further reduce the risk associated with glide avalanches by improving forecasting and control methods. Acknowledgments Thanks to the many people who shared their ideas on glide avalanches, particularly Bruce Jamieson, Doug Wilson, Johann Slam and Tony Moore. Thanks to Bruce Jamieson and Ron Perla for providing photographs of glide avalanches for use in this paper. 59

9 References Akitaya, E. and Y. Endo Studies of the behavior of a snow cover on slope XVIII. Glide motion of snow and formation of crack in melting season. Low Temperature Science, Ser. A, 43, Bader, H., R. Haefeli, E. Bucher, I. Neher, O. Eckel and Chr. Thams Der Schnee und seine Metamorphose (Snow and its Metamorphism). Beitrage zur Geologie der Schweiz, Geotechnische Serie, Hydrologie, Lieferung 3, Bern [English Translation by Snow, Ice Permafrost Research Establishment, Corps of Engineers, U.S. Army. Translation 14, January 1954], 313 pp. Clark, J. and D.M. McClung Full-depth avalanche occurrences caused by snow gliding, Coquihalla, British Columbia, Canada. Journal of Glaciology, 45(150), Endo, Y Glide processes of a snow cover as a release mechanism of an avalanche on a slope covered with bamboo bushes. Contributions to the Institute of Low Temperature Science, Hokkaido University Series A, 32, Endo, Y Release mechanism of an avalanche on a slope covered with bamboo bushes. Annals of Glaciology, 6, Endo, Y. and E. Akitaya Glide mechanism of a snow cover on a slope covered with dwarf bamboo bushes. Deuxieme Rencontre internationale sur la neige et les avalanches, et 14 avril, 1978, Grenoble, France. Association nationale pour l etude de la neige et des avalanches, Grenoble, France, In der Gand, H.R. and M. Zupanièiè Snow gliding and avalanches. International Association of Scientific Hydrology Publication 69 (Symposium at Davos 1965 Scientific Aspects of Snow and Ice Avalanches), Lackinger, B Stability and fracture of the snow pack for glide avalanches. International Association of Hydrological Sciences Publication 162 (Symposium at Davos 1986 Avalanche Formation, Movement and Effects, Lang, T.E Wave pattern of flowing snow slabs. Journal of Glaciology, 19(81), McClung, D.M Creep and glide processes in mountain snowpacks. National Hydrology Research Institute (NHRI) Paper No. 6, Ottawa, 66 pp. McClung, D.M A physical theory of snow gliding. Canadian Geotechnical Journal, 18(1), McClung, D.M Mechanics of snow slab failure from a geotechnical perspective. International Association of Hydrological Sciences Publication 162 (Symposium at Davos 1986 Avalanche Formation, Movement and Effects), McClung, D.M. and G.K.C. Clarke The effects of free water on snow gliding. Journal of Geophysical Research, 92(7), McClung, D.M and P.A. Schaerer The Avalanche Handbook. Seattle, WA, The Mountaineers, 271 pp. McClung, D.M., S. Walker and W. Golley Characteristics of snow gliding on rock. Annals of Glaciology, 19, Rice, B., D. Howlett and R. Decker Preliminary investigations of glide/creep motion sensors in Alta, Utah. In ISSW 96, International Snow Science Workshop, Banff, Canada, Seligman, G Snow Structure and Ski Fields. International Glaciological Society, Cambridge, England. Wilson, A., G. Statham, R. Bilak and B. Allen Glide avalanche forecasting. In ISSW 96, International Snow Science Workshop, Banff, Canada,